Biochemistry 1985, 24, 1928-1 933
1928
bow, M. S., Ed.) pp 441-455, Alan R. Liss, New York. Gall, C. M., Cross, T. A., DiVerdi, J. A., & Opella, S. J. (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 101-105. Hagen, D. S., Weiner, J. H., & Sykes, B. D. (1978) Biochemistry 17, 3860-3866. Hagen, D. S., Weiner, J. H., & Sykes, B. D. (1979) Biochemistry 18, 2007-2012. Incardona, N. L., & Kaesberg, P. (1964) Biophys. J. 4, 11-21. Jost, P., Griffith, 0. H., Capaldi, R. A,, & Vanderkooi, G. (1973) Proc. Natl. Acad. Sci. U.S.A. 70, 480-484. Kehry, M. R., Wilson, M. L., & Dahlquist, F. W. (1983) Anal. Biochem. 131, 236-241. Lu, P., Jarema, M., Mosser, K., & Daniel, W. E. (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 3471-3475. Makino, S., Woolford, J. L., Tanford, C . , & Webster, R. E. (1975) J . Biol. Chem. 250, 4327-4332. Mazer, N. A., Benedek, G. B., & Carey, M. C. (1980) Biochemistry 19, 601-615. Muller, K. (1 98 1 ) Biochemistry 20, 404-4 14. Nakashima, Y., & Konigsberg, W. (1 974) J . Mol. Biol. 88, 598-600. Nozaki, Y., Reynolds, J. A., & Tanford, C. (1978) Biochemistry 17, 1239-1246.
Oldfield, E., Gilmore, R., Glaser, M., Gutowsky, M. S., Ksung, J. C., Kang, S. Y., King, T. E., Meadows, M., & Rice, D. (1978) Proc. Natl. Acad. Sci. U.S.A. 75, 4657-4660. Opella, S. J., Cross, T. A., DiVerdi, J. A., & Sturm, C. F. (1980) Biophys. J . 32, 531-548. Paddy, M., Dahlquist, F. W., Davis, J. H., & Bloom, M. (1981) Biochemistry 20, 5755-5759. Robinson, I. K., & Harrison, S. C. (1982) Nature (London) 297, 563-568. Smilowitz, H., Carson, J., & Robbins, P. W. (1972) J . Supramol. Struct. 1 , 8-18. Sykes, B. D., & Weiner, J. H. (1980) in Magnetic Resonance in Biology (Cohen, J. S . , Ed.) Vol. 1 , pp 171-196, WileyInterscience, New York. Turnell, D. C., & Cooper, J. D. H. (1982) Clin. Chem. ( Winston-Salem, N.C.) 28, 521-53 1 . Van Wezenbeek, P., Hulsebos, T., & Schoenmakers, J. (1980) Gene 1 1 , 129-148. Vik, S., & Capaldi, R. A. (1977) Biochemistry 16,5755-5759. Wickner, W. (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 1159-1163. Woolford, J. L., & Webster, R. E. (1975) J. Biol. Chem. 250, 43 33-43 39.
Conformational Flexibility of Neurophysin As Investigated by Local Motions of Fluorophores. Relationships with Neurohypophyseal Hormone Binding? Mohamed Rholamt*i and Pierre Nicolas*,s Department of Biochemistry, School of Chemical Sciences, University of Illinois, Urbana, Illinois 61801, and Groupe de Neurobiochimie Cellulaire et Moldculaire, Uniuersitd Pierre et Marie Curie, 75006 Paris, France Received August 14, 1984
Flexibility of various structural domains of neurophysin and neurophysin-neurohypophyseal hormone complexes has been investigated through the fast rotational motion of fluorophores in highly viscous medium. Despite seven intrachain disulfide links, it is shown that some domains of neurophysin remain highly flexible. Dimerization of neurophysin does not affect the structural integrity of the individual subunits, each subdomain being conformationally equivalent within each protomer of the unliganded dimer. The absence of heterogeneous fluorescence anisotropy precludes the existence of a dimer tautomerization equilibrium. Binding of the hormonal ligands to neurophysin dimer promotes a large conformational change over the whole protein structure as assessed by differential alterations of the flexibility-rigidity and intrasegmental interaction properties of domains that do not participate directly to the dimerization/binding areas. The order of free-energy coupling between ligand binding and protein subunit association has been evaluated. Data are consistent with a model in which the first mole of bound ligand stabilizes the dimer by increasing the intersubunit contacts while the second mole of ligand induces most of the described conformational change. Accordingly, the positive cooperativity between the two dimeric binding sites is linked mainly to the binding of the second ligand. The induced structural change is perceived differently by each subunit as assessed by opposite local motions of Tyr49in each liganded protomer and leads to the formation of a dimeric complex with a global pseudospherical symmetry although containing domains of local asymmetry. ABSTRACT:
T e posterior pituitary contains a class of highly disulfidelinked proteins, neurophysins, associated with both biosynthesis and transport of the neurohypophyseal hormones, ocytocin and This work was supported in part by funds from the Universiti Pierre et Marie Curie, the CNRS (Unit6 AssociCe 554), and INSERM (CRE 834006). *Department of Biochemistry, University of Illinois. 5 Groupe de Neurobiochimie Cellulaire et Moliculaire, UniversitB Pierre et Marie Curie.
0006-2960/85/0424-1928$01.50/0
vasopressin, along the hypothalamo neurohypophyseal tract [for recent reviews, see Cohen et al. (1983), Chaiken et al. (1983), Richter (1983), Breslow (1984), and Pickering & Swann (1984)l. Within the neurosecretory granule two distinct neurophysins are found in noncovalent association with a different hormone although each neurophysin is able to bind vasopressin or ocytocin with similar affinities in vitro. Studies of the polymerization process of neurophysins demonstrated that these proteins self-associate in a form of 0 1985 American Chemical Society
CONFORMATIONAL FLEXIBILITY OF NEUROPHYSIN
a dimer in aqueous solutions (Nicolas et al., 1976, 1980; Pearlmutter, 1979). Both the insensitivity to pH and temperature of the dimerization equilibrium and kinetic measurements suggested that protomer association areas are mainly hydrophobic in nature (Pearlmutter, 1979; Nicolas et al., 1980). Hydrodynamic parameters of the monomer and the dimer favored an entropy-driven side-by-side association model allowing maximum contact between subunits and stabilization of the dimer (prolate ellipsoid of axial ratio a / b = 3.6) through numerous weak interactive forces (Rholam & Nicolas, 1981). Dimer formation seems to be achieved without profound conformational rearrangement in the monomer structure as assessed by spectroscopical evidence (Nicolas et al., 1978a, 1979; Sur et al., 1979). Noncovalent complexes formed between neurophysins and hormones, or peptide analogues, have been submited to spectroscopical and hydrodynamic analysis. Facilitated dimerization is observed under conditions of preferential binding to the dimeric form of hormones and analogues (Nicolas et al., 1976, 1978a, 1980; Pearlmutter & McMains, 1977; Pearlmutter & Dalton, 1980) and positive binding cooperativity has been detected between the two strong dimeric sites common to both hormones (Hope et al., 1975; Nicolas,et al., 1978a,b; Pearlmutter & Dalton, 1980; Tellman & Winzor, 1980). Spectroscopic studies have provided some evidence that complex formation should produce local rearrangements in the neurophysin tertiary structure. In particular, UV absorption spectra arising from the single tyrosine side chain in position 49 indicate that this moiety is displaced from a hydrophobic environment to a more polar one upon ligand binding (Griffin et al., 1973; Wolff et al., 1975). CD spectroscopy additionally indicated slight perturbation arising from neurophysin disulfide links (Breslow & Weiss, 1972). Recent hydrodynamic studies demonstrated that ligand binding leads to the formation of compact pseudospherical dimeric complexes (Rholam et al., 1982). Such a ligand-induced conformational change was postulated to be linked with the intradimeric cooperativity properties of the dimer. The real extent to which the tertiary structure of such a small, highly constrained, protein is flexible and is able then to accommodate large conformational perturbations is unknown. The above findings indicate that neurophysin tertiary structure has indeed some flexibility and undergoes more profound and extented changes upon binding than those detected through the limited Tyr49and disulfide bridge spectroscopical probes. In order to gain relevant and general information on the flexibility-rigidity properties of that particular class of proteins and to detect the structural changes induced by polymerization and/or binding, changes in local environmental flexibility of various subdomains of neurophysins and neurophysin-hormone complexes have been investigated through the modifications in the fast rotational motions of fluorophore residues in highly viscous medium following the new approach developed by Weber et al. (1984). MATERIALS AND METHODS Highly purified neurophysins I and I1 were prepared by isoelectric focusing (Camier et al., 1973). Mononitrated Tyr49 neurophysins were obtained by reaction with tetranitromethane and purified by electrophoresis as previously reported (Wolff et al., 1975). The dansyl (Dns)fluorescent group was covalently attached to the protein through residues Ala1 and Lysl* (or LysS9)as previously reported (Rholam & Nicolas, 1981). The tripeptides Cys(S-Me)-Tyr-Ile-NH, and Cys(SMe)-Phe-Ile-NH, were from Bachem. Their spectroscopic purity was assessed by absorption and fluorescence measure-
VOL. 24, NO. 8, 1985
1929
ments. Spectral-quality glycerol was from Aldrich. Fluorescence polarization measurements were done with the apparatus described by Jameson et al. (1976). The values were corrected for solvent background if its contribution to the total intensity exceeded 0.5%. Fluorescence lifetimes were determined by the cross-correlation phase method (Spencer & Weber, 1969) with updated electronics from SLM (SLM Instruments, Urbana, IL). Corning 0-53 filters were used on tyrosyl emission, and the exciting light was isolated with Beckman monochromator set at 280 nm and a 7-54 Corning filter. For dansyl fluorescence, the monochromator was set at 350 nm, and Corning 3-73 filters were used for emission. Temperature was regulated by a methanol circulating bath, and optical modules were purged with dry nitrogen to prevent frosting. Viscosity values for glycerol-water mixtures were from Miner & Dalton (1953). Native and dansylated neurophysins were dissolved in 80% glycerol-20% phosphate buffer (0.05 M, pH 7.0). The polarization and lifetime of the solutions were determined in the interval -40 to 20 "C. The data were analyzed by employing the relation established in a previous paper (Weber et al., 1984): Y = In [ A ( O ) / A ( T )- 11 - In [ R T ( 7 ) / V ]= - In ~ ( 0 ) b(T - To)
+
where A(0) is the limiting anisotropy, A( T ) is the value of the anisotropy at the Kelvin temperature T, R is the gas constant, ( 7 ) is the mean lifetime of the excited state, Vis the effective volume of the fluorophore, ~ ( 0 is ) the viscosity at the temperature To and b is the thermal coefficient of the "local" viscosity that characterizes the resistance that the environment offers to the rotation of fluorophores. In the general case (Rholam et al., 1984; Scarlata et al., 1984), the plot of Y against the temperature t = ( T - To)yields two slopes corresponding to two regions in which the rotational motions of the fluorescent residue are limited either by the external solvent [b(S)]or by the peptide environment [b(U)], respectively. RESULTS Local Motions in Neurophysins. The Y plot of native neurophysin against the temperature t is presented in Figure 1. This plot is characterized by two linear parts corresponding to two different regimes (Rholam et al., 1984; Scarlata et al., 1984): one at low temperature whose slope is equal to the thermal coefficient of the solvent [ b ( S ) = 7%] and the other one at high temperature whose thermal coefficient is much smaller [b(U) = 3.2%]. The transition from one region to the other occurs at a critical temperature (tc = -12 "C) where the motions of the tyrosyl groups become restricted by the protein environment. As it is shown in Figure 2, a similar pattern is obtained for dansylated neurophysin except that two high-temperature slopes are observed: (b(U)), = 4.3% and (b(U)), = 5.9%. This result was expected since each subunit has two distinct sites for dansyl labeling, Ala' and Lys'* (or L Y S ~(Rholam ~) & Nicolas, 1981). On the basis of the values of both the apparent thermal coefficients, (b(U)), and the relative contributions of each fluorophore to total fluorescence emission, Ai) (Rholam & Nicolas, 198l), the values of b,(U) have been calculated by the relation previously reported (Scarlata et al., 1984). The same data (Table IA) have been obtained for both native and dansylated neurophysin, under conditions in which either the monomer or the dimer is the prevalent species (Nicolas et al., 1980). This finding agrees with the results
1930 B O C H E M I S T R Y
RHOLAM A N D NICOLAS
I Table I: Neurophysin Parameters Obtained in the Presence or Absence of Peptide Analogues 100(b(U)) 100b(U) t, amplitude species ("C-I) ("C-I) ("C) (deg)" (A) Neurophysins neurophysin 3.2 -12 13 neurophysin-Dnsb 4.3 4.9 (i) 8 26 13 5.9 2.2 (ii) -12
(B) Neurophysin Complexes
N02-Tyr49-Tyr2 T~r~~-Phe~
5.5
10 5 -26 8 -26 -3
3.4 5.35 4.5 5.4
T~r~~-Tyr~
28 23 I 26 I 18
neurophysin-Dns-Tyr2 4.1 "The amplitude was calculated from the relation determined i n a previous paper (Rholam et al., 1984): (cos = [ l + 2A(t,)/ A ( 0 ) ] / 3 . bTheparameter values of each fluorophore, (i) Dns-Ala' and (ii) Dns-LysIsor Dns-Ly~~~, were determined from the relation (Scarlata et al., 1984) (b(U)) = m i ) b , ( U ) . -40
-20
t
0
20
FIGURE 1: Plot of Y vs. temperature t of the native neurophysin dissolved in 80% glycerol-20% 0.1 M acetate buffer, pH 6.2.
-1.4-
-1.8-
- 2.2-
- 2.6-3
- 40
- 20
0
1
20
-40
t FIGURE 2: Plot of Y vs. temperature t of the dansylated neurophysin under the same conditions as given in Figure 1.
obtained by other methods (Pearlmutter & Dalton, 1980; Rholam & Nicolas, 1981) and indicates that tyrosyl and dansyl immediate environments are unperturbed upon dimerization. Ligand Effects on Local Motions. Since neurophysins contain a strong site per monomer that binds hormones or peptide analogues (Nicolas et al., 1978a), two N-terminal tripeptide analogues of ocytocin have been used to analyze the rotational motions of the tyrosyl groups of neurophysins in the following dimeric complexes:' (A) N02-Tyr49-TyrZ; (B) T ~ r ~ ~ - P (C) h e ~T;~ r ~ ~ - T yAsr ~the . fluorescence of the tyrosine in neurophysin can be completely quenched by nitration (N02-Tyr49),the motions of Tyr2 of the bound substrate can be examined by following the fluorescence depolarization of I Emission from any nonbound ligand or exposed tyrosyl groups was quenched by addition of a 100-fold molar excess of sodium citrate (final concentration 0.1 M).
-
-x)
-20
-10
0
IO
20
30
t FIGURE 3: Y plot of the neurophysin-Cys(S-Me)-Phe-Ile-NH2complex. The experimental conditions are the same as those in Figure
1.
Cy@-Me)-Tyr-Ile-NHz liganded to nitrated neurophysin dimer (complex A). On the other hand, the motions of Tyr49 in the neurophysin were analyzed through the fluorescence depolarization of the native protein complexed with the tripeptide Cys(S-Me)-Phe-Ile-NH2(complex B). Complex C allowed the observation of motions of tyrosine of both protein and substrate. These experiments were conducted in the presence of saturating amounts of tripeptides (ligand to protein molar ratio = 10). One high-temperature slope is observed for complex A (Table IB) whereas two high-temperature slopes characterize complex B (Table IB and Figure 3). Hence, we expected to observe three high-temperature slopes for complex C corresponding to a linear combination of those obtained in complexes A and B. However, due to very close critical temperature values of Tyr2 and one of the Tyr49residues in the dimer
CONFORMATIONAL FLEXIBILITY OF NEUROPHYSIN
VOL. 24,
NO.
8, 1985
1931
Table 11: Free Energies of Ligand Binding and Dimerization Process of Neurophysin at T = 298 K association constants free energies equilibrium (mol-I)' (kcal mol-') M+L*ML kl = 6.11 x 104 AGO11 = -6.53 D+L*DL k2 = 1.3 X lo5 AGO21 = -6.98 DL + L DL2 k'z = 5.35 X lo5 AGOz2 = -7.82 'The constant values were taken from Nicolas et al. (1978b).
which induces a maximum conformational rearrangement in the structure of neurophysin dimer (Rholam et al., 1982).
FIGURE 4: Y plot of the neurophysin-Cys(S-Me)-Tyr-Ile-NH2complex. The conditions used are the same as those in Figure 1.
(Scarlata et al., 1984), only two thermal coefficient values of (b(U)) have been determined (Table IB and Figure 4). These observations indicate that the Tyr2 groups are located in two identical domains that correspond to the high-affinity dimeric sites. Moreover, the ligands act differently on the conformation of neurophysins so that the dimeric environments of Tyr49 groups (one per monomer) become unequivalent. In contrast, the analysis of local motions of the dansyl residues of neurophysin indicates that, upon ligand binding, only one value of the parameters, t, and b(U), can be determined (Table IB). Thus, upon ligand binding the dansyl groups move into environments that are not experimentally distinguishable. Also, the hormonal ligands produce the same effect in each protomer of the liganded dimer. Order of Free-Energy Couplings. While the binding of peptide ligands to neurophysins is cooperative (Nicolas et al., 1978a,b) and promotes pseudospherical complexes (Rholam et al., 1982), the mechanism by which this intradimeric cooperativity occurs is still unknown. In view of the importance of the relation between this phenomenon and the above data, the order of free-energy couplings between ligand binding and protein subunit association has been analyzed by the following equation (Weber, 1970, 1972, 1984): = (AGzz - AGii)/(AGzi - 4 1 ) where AGll designates the standard free energy of ligand binding to the monomer while AGzl and AGzz are respectively the standard free energies of associationof the first and second molecule of ligand L to the dimer. If Z = 1, the free-energy couplings are of intermediate order. If Z >> 1, they are practically of second order, and if Z
